Methods Applying AHRQ Quality Indicators to Healthcare ... - HCUP

HCUP Methods Series

patterns, patient safety, access to health care programs, and outcomes of treatments at thenational, State and local market levels.Two HCUP discharge datasets were used for the NHQR: The HCUP Nationwide Inpatient Sample (NIS), a nationally stratified sample ofhospitals (with all of their discharges) from States that contribute data to the NISdataset (42 States in the 2008 NIS). The HCUP State Inpatient Databases (SID), a census of hospitals (with all of theirdischarges) from 42 participating States in 2008.For 2008, the NIS contains roughly 8.2 million discharges from more than 1,000 hospitals andthe SID contains about 35.6 million discharges (approximately 89 percent of the 39.9 milliondischarges in the United States). Data from the 1994, 1997, and 2000-2008 NIS were used fornational estimates of QI rates in the NHQR. Data from the 2000, 2004, 2007, and 2008 SIDwere used for State-level estimates, for States that agreed to participate. Data from these samedata years were also used for reporting by State-level subpopulations (race/ethnicity,community income quartile, and expected primary payer). For the list of data organizations thatcontribute to the HCUP databases, see Table 2 at the end of this methods report.In preparation for the NHQR, NHDR, and derivative products, the HCUP databases needed tobe customized as indicated below:1. The HCUP SID were modified to create analytic files consistent across States.Subset to Community Hospitals. For the SID, we selected community 1 hospitals andeliminated rehabilitation hospitals.Weight for Missing Hospitals. Because some statewide data organizations do not reportdata for all community hospitals in the State, we weighted hospitals in the SID to theState’s universe of hospitals in the American Hospital Association (AHA) Annual SurveyDatabase based on hospital characteristics. Weight for Missing Quarters. Discharges from hospitals operating for the entire year butnot contributing data for one or more quarters were weighted up to annual estimates forthat institution in the SID.2. The NIS and SID were augmented as necessary for the NHQR and NHDR analyses:Impute for Missing Characteristics. For missing age, gender, race/ethnicity, ZIP Code,and expected primary payer data that occurred on a small proportion of dischargerecords, we used a “hot deck” imputation method (which draws donors from strata ofsimilar hospitals and patients) to assign values while preserving the variance within thedata.1 Community hospitals are defined by the AHA as “non-Federal, short-term, general, and other specialtyhospitals, excluding hospital units of institutions.” The specialty hospitals included in the AHA definition of“community hospitals” are: obstetrics-gynecology, ear-nose-throat, short-term rehabilitation, orthopedic,and pediatric institutions. The AHA also groups public hospitals and academic medical centers withcommunity hospitals. Excluded from the AHA definition of “community hospitals” are long-term hospitals,psychiatric hospitals, and alcoholism/chemical dependency treatment facilities. For the NHQR analyses,we select all AHA-defined “community hospitals” with the exception of short-term rehabilitation hospitals(beginning with 1998 HCUP data).Methods for HCUP Data in 2011 NHQR and NHDR 2December 9, 2011

Assign Additional Measures for Reporting. We assigned median household incomequartile using the Claritas ZIP Code data linked to patient’s ZIP Code in the SID.Beginning with the 2008 NHQR and NHDR, we added reporting by the National Centerfor Health Statistics (NCHS) county-level classification of urban-rural location, whichincludes gradations of metropolitan, micropolitan, and non-core counties by populationsize.3. For the NHDR, the HCUP SID were used to create disparities analysis files designed toprovide national- and State-level estimates for the report and derivative products. Of the 42States participating in the 2008 SID, the following 31 HCUP States report race/ethnicity ofdischarges: Arizona, Arkansas, California, Colorado, Connecticut, Florida, Georgia, Hawaii,Kansas, Kentucky, Maine, Maryland, Massachusetts, Michigan, Missouri, Nevada, NewHampshire, New Jersey, New York, Oklahoma, Oregon, Pennsylvania, Rhode Island, SouthCarolina, Tennessee, Texas, Utah, Vermont, Virginia, Wisconsin, and Wyoming.The SID were used to create a disparities analysis file that was designed to provide nationalestimates for the NHDR, using a weighted sample of hospitals from the 31 HCUP States.Appendix A to this report provides detail on the creation of the disparities analysis file fornational estimates. The individual SID were used to create additional disparities analysisfiles for State-level reporting by race/ethnicity. Appendix B to this report provides detail onthe creation of disparities analysis files for State-level estimates. For the list of dataorganizations that contribute race and ethnicity data to the HCUP databases (and that wereincluded in the disparities analysis file), see Table 2 at the end of this methods report.4. The SID were also used for reporting by State and State-level subpopulations(race/ethnicity, community income quartile, and expected primary payer). Given the varieddistribution of race, ethnicity, and socioeconomic groups across states, policymakersincreasingly want to know if and how quality of care varies for these different populations.State-level QI estimates are only reported for participating HCUP Partners that agree torelease information.STEPS TAKEN TO APPLY AHRQ QUALITY INDICATORS TO THE HCUP DATATo apply the AHRQ Quality Indicators to HCUP hospital discharge data for the NHQR andNHDR, several steps were taken: (1) QI software review and modification, (2) acquisition ofpopulation-based data, (3) assignment of QIs to the HCUP databases, and (4) identification ofstatistical methods.1. Review and Modify QI Software. For the 2011 NHQR and NHDR, we used a “modifiedversion” of the 4.1 software. We started with version 4.1a, included software correctionsfrom version 4.1b, then added software corrections (but not definitional changes) fromversion 4.2. In addition, we did not utilize the present on admission (POA) estimationmodule for the IQIs, PDIs, and PSIs since POA indicators were not uniformly available fromStates that contribute to the HCUP databases. Specific modifications are noted as footnotein the tables. Because each of these software modules was developed for State andhospital-level rates, rather than national rates, some changes to the QI calculations werenecessary.We added four indicators particularly relevant to the structure of the NHQR and NHDR. Twoindicators were created for discharges age 65 years and older: immunization-preventableinfluenza, age 65 and over; and asthma admissions, age 65 and over. Two additionalMethods for HCUP Data in 2011 NHQR and NHDR 3December 9, 2011

indicators were created from a modified version of the chronic and overall PQI compositesthat excluded chronic obstructive pulmonary disease to facilitate longitudinal analyses.Because of ICD-9-CM coding changes, chronic obstructive pulmonary disease estimates(PQI 05) for data prior to 2005 are not compatible with rates for 2005 forward.2. Acquire Population-Based Data for Denominators and Risk-Adjustment. The next stepwas to acquire data for the numerator and denominator populations for the QIs. A QI is ameasure of an event that occurs in a hospital, requiring a numerator count of the event ofinterest and a denominator count of the population (within a hospital or geographic area) towhich the event relates.For the numerator counts of the AHRQ QIs, we used the HCUP NIS or NHDR disparitiesanalysis file to create national estimates and used the SID for State-level estimates. For thedenominator counts, we identified two sources for all reporting categories and for alladjustment categories listed in the HCUP-based tables. For QIs that related to providers,the HCUP data were used for State- and national-level discharge denominator counts. ForQIs that related to geographic areas, population ZIP-Code-level counts from Claritas (avendor that compiles and adds value to the U.S. Bureau of Census data) were used fordenominator counts. Claritas uses intra-census methods to estimate household anddemographic statistics for geographic areas (Claritas, Inc., 2008). We also used the Claritaspopulation data for risk adjustment by age and gender for the area-based QIs.3. Assign QI Indicators to the HCUP Databases. The four AHRQ QI program modules wereapplied to the prepared SID data using all available diagnoses and procedures reported byeach State. The QI indicators from the SID were then linked to the corresponding dischargerecords on the NIS.4. Adapt Statistical Methods to HCUP Data. Several statistical issues needed to beaddressed when applying the AHRQ QI software to the HCUP data, including: age-genderadjustment for all QIs; severity/comorbidity adjustment for the discharge-based IQIs, PSIs,and PDIs; and derivation of standard errors and appropriate hypothesis tests.Age-Gender Risk Adjustment. For the PQIs and area-based IQIs, PSIs, and PDIs, theobserved rates were risk-adjusted for age and gender differences across populationsubgroups and were based on methods of direct standardization (Fleiss, 1973). Agewas categorized into 18 five-year increments (described in Table 3, Age Groupings forRisk Adjustment). Although the AHRQ QI software uses a similar approach to adjust thearea-based QIs, we relied on direct standardization because of the additional reportingcategories and population denominators required in the NHQR.Age, Gender, Severity, and Comorbidity Risk Adjustment.For the discharge-based PSIs, the observed rates were risk-adjusted for age, gender,age-gender interaction, DRG cluster, and comorbidity using the regression-basedstandardization that is part of the AHRQ PSI software, with the following exceptions.When reporting by age, the risk adjustment includes all of the above except age. Whenreporting by gender, the risk adjustment includes all of the above except gender.For the discharge-based IQIs, risk adjustments were made for age, gender, age-genderinteraction, and 3M All Patient Refined Diagnosis Related Groups (APR-DRGs) risk ofmortality or severity score using the regression-based standardization that is part of theAHRQ IQI software, with the following exceptions. When reporting by age, the riskadjustment includes all of the above except age. When reporting by gender, the riskadjustment includes all of the above except gender.Methods for HCUP Data in 2011 NHQR and NHDR 4December 9, 2011

For the discharge-based PDIs, risk adjustments were made for age, gender, DRG andMDC clusters, and comorbidity using the regression-based standardization that is part ofthe AHRQ PDI software. Measure-specific stratification by risk group, clinical category,and procedure type was also applied, with the following exceptions. When reporting byage, the risk adjustment includes all of the above except age. When reporting bygender, the risk adjustment includes all of the above except gender.Standard Errors and Hypothesis Tests. Standard error calculations for the rates werebased on the HCUP report entitled Calculating Nationwide Inpatient Sample (NIS)Variances (Houchens, et al., 2005). There is no sampling error associated with Claritascensus population counts; therefore, appropriate statistics were obtained through theStatistical Analysis System (SAS) procedure called PROC SURVEYMEANS.Masking Rates for Statistical Reliability, Data Quality, and Confidentiality. QI estimateswere included in the NHQR and NHDR if they reached a threshold defined by a relativestandard error less than 30% and at least 11 unweighted cases in the denominator.Estimates that did not satisfy these criteria were set to missing. Statistical calculationsare explained in Appendix C to this report.SPECIAL ANALYSESCalculating Costs Associated with Quality IndicatorsThe NHQR includes trends in total national costs from 2000 to 2008 for the three PQI compositemeasures – for overall, acute, and chronic conditions. Total national costs associated withpotentially avoidable hospitalizations were calculated overall for the U.S., by income quartile,and by race/ethnicity.Total charges were converted to costs using the hospital-level HCUP Cost-to-Charge Ratiosbased on Hospital Cost Report data from the Centers for Medicare and Medicaid Services(CMS). 2 Costs reflect the actual costs of production, while charges represent what the hospitalbilled for the stay. Hospital charges reflect the amount the hospital charged for the entirehospital stay and do not include professional (physician) fees. The total cost is the product ofthe number of stays for each QI measure and the mean cost for each QI measure. Thisapproach compensates for stays for which charges (and thus estimated costs) are not available.Costs were adjusted to 2008 dollars for all years using the price indexes for the gross domesticproduct (downloaded from the Bureau of Economic Analysis, U.S. Department of Commerce,June 1, 2010).Calculating IQI and PSI Summary MeasuresTo examine national and state-level trends in inpatient mortality and patient safety events, riskadjustedrates for select Inpatient Quality Indicators (IQIs) and Patient Safety Indicators (PSIs)were summarized. The three NHQR/NHDR summary measures include: (1) Mortality forselected conditions based on select IQIs; (2) Mortality for selected procedures based on selectIQIs; and (3) Patient Safety based on select PSIs. These summary measures were calculatedas a weighted sum of risk-adjusted rates for individual IQIs and PSIs. Additional information onthe calculation of IQI and PSI Summary Measures is provided in Appendix D.2 HCUP Cost-to-Charge Ratio Files. Healthcare Cost and Utilization Project (HCUP). August 2011.Agency for Healthcare Research and Quality, Rockville, MD. www.hcupus.ahrq.gov/db/state/costtocharge.jsp.Methods for HCUP Data in 2011 NHQR and NHDR 5December 9, 2011

Determining Benchmarks for State Performance for the Quality IndicatorsBased on a recommendation from the Institute of Medicine’s report on Future Directions for theNational Healthcare Quality and Disparities Reports, benchmarks based on a straight averageof the top 10 percent of reporting States were determined. For a benchmark to be calculated,rates for at least 30 States needed to be available. Top performing States are not identified.Comparison of Inpatient and Emergency Department Use for Selected PQIs and PDIsBeginning in the 2009 NHQR, the HCUP Nationwide Emergency Department Sample (NEDS)and NIS data were used to examine national and regional differences in inpatient andemergency department use for selected PQIs and PDIs. Details for this analysis are provided inAppendix E.Readmissions for Select Chronic ConditionsHCUP data was used to examine differences in hospital readmissions for congestive heartdisease, asthma, asthma with chronic obstructive pulmonary disease, pneumonia, and acutemyocardial infarction across States, income quartile, age, gender, and race/ethnicity. Details forthis analysis are provided in Appendix F.CAVEATSSome caution should be used in interpreting the AHRQ QI statistics presented in the NHQR andNHDR. These caveats relate to the how the QIs were applied, ICD-9-CM coding changes, inter-State differences in data collection, and other more general issues.Rehabilitation Hospitals: These hospitals are excluded from the 2000-2008 NIS but includedin the 1994 and 1997 NIS because of the change in the NIS sampling strategy (beginning in the1998 NIS). Patients treated in rehabilitation hospitals tend to have lower mortality rates andlonger lengths of stay than patients in other community hospitals, and the completeness ofreporting for rehabilitation hospitals is very uneven across the States. The elimination ofrehabilitation hospitals in 2000-2008 may affect trends in the QIs; however, based on previousanalyses, the effect is likely small since only 3 percent of community hospitals are involved.ICD-9-CM Coding Changes: A number of the AHRQ QIs are based on diagnoses andprocedures for which ICD-9-CM coding has generally become more specific over the period ofthis study. If coding changes cause earlier estimates to not be comparable to the laterestimates, then the earlier estimates are not reported. For this reason, the PQI for chronicobstructive pulmonary disease (PQI 5), the overall PQI composite (PQI 90), and chronic PQIcomposite (PQI 92) are not reported prior to 2005. In addition, birth trauma (PSI 17) is notreported prior to 2004, and QIs for postoperative hemorrhage (PSI 9 and PDI 8) are notreported before 1997.Data Collection Differences Among States: Organizations providing statewide data generallycollect the data using the Uniform Billing format (UB-92 or UB-04) and, for earlier years, theUniform Hospital Discharge Data Set (UHDDS) format. However, not every statewide dataorganization collects all data elements nor codes them the same way. For the NHQR andNHDR, uneven availability of a few data elements underlie some estimates, as noted next.Methods for HCUP Data in 2011 NHQR and NHDR 6December 9, 2011

Data Elements for Exclusions: Three data elements required for certain QIs were notavailable in every State: “secondary procedure day,” “admission type” (elective, urgent,newborn, and emergency), and “present on admission.” We modified the AHRQ QI software ininstances where these data elements are used to exclude specific cases from the QI measures:The PSIs and PDIs that use secondary procedure day were modified to calculateindicators without considering the timing of procedures. Affected PSIs and PDIs areshown in Table 4.For QIs that use admission type “elective” and “newborn,” we imputed the missingadmission type using available information. For all States except California, anadmission type of “elective” was assigned if the DRG did not indicate trauma,delivery, or newborn. An admission type of “newborn” was assigned if the DRGindicated a newborn. For California, which did not provide any information onadmission type, information on scheduled admissions was used to identify electiveadmissions and DRGs were used to identify newborn admissions.For QIs that use present on admission (POA), we modified the AHRQ QI software tocalculate indicators without considering whether the condition was present atadmission. PSIs and PDIs that use POA are shown in Table 5.Number of Clinical Fields: Another data collection issue relates to the number of fields thatstatewide data organizations permit for reporting patients’ diagnoses and procedures during thehospitalization. The SID for different States generally contain as few as 6 or as many as 30fields for reporting diagnoses and procedures, as shown in Table 6 at the end of this methodsreport. The more fields used, the more quality-related events that can be captured in thestatewide databases. However, in an earlier analysis, even for States with 30 diagnosis fieldsavailable in the year 2000, 95 percent of their discharge records captured all of patients’diagnoses in 10 to 13 data elements. For States with 30 procedure fields available, 95 percentof records captured all of patients’ procedures in 5 fields. Thus, limited numbers of fieldsavailable for reporting diagnoses and procedures are unlikely to have much effect on results,because all statewide data organizations participating in HCUP allow at least 9 diagnoses and 6procedures. We decided not to artificially truncate the diagnosis and procedure fields used forthe NHQR analyses, so that the full richness of the databases would be used.E Codes: Another issue relates to external cause-of-injury reporting. Eight of the 27 PatientSafety Indicators and three of the Pediatric Quality Indicators use E code data to help identifycomplications of care or to exclude cases (e.g., poisonings, self-inflicted injury, trauma) fromnumerators and denominators, as shown in Table 7 at the end of this methods report. AlthoughE codes in the AHRQ PSI and PDI software have been augmented wherever possible with therelated non-E codes in the ICD-9-CM system, E codes are still included in some AHRQ PSI andPDI definitions. Uneven capture of these data has the potential of affecting rates and should bekept in mind when judging the level of these events. While all HCUP States report E Codes, thepolicies on reporting medical misadventures and adverse effects can vary. In particular,California and Washington do not require hospitals to report E codes in the range E870-E879(medical misadventures and abnormal reactions). Georgia does not report E codes in the rangeE870-E879 (medical misadventures and abnormal reactions) and E930-E949 (adverse effects).SC does not report E codes in the range E870-E876 (medical misadventures). West Virginiadoes not require hospitals to report any E Codes.Effects of Adding New States to the NIS: Over time, HCUP has expanded through theparticipation of additional statewide data organizations. Because each NIS is a sample ofhospitals from the States participating in that year (and weighted to the universe of communityMethods for HCUP Data in 2011 NHQR and NHDR 7December 9, 2011

hospitals nationally), potential exists for different practice patterns across States to influencenational measures related to clinical practice over time.The table below lists the States that were added to HCUP between the years used in this report.Period States1994 AZ, CA, CO, CT, FL, IL, IA, KS, MD, MA, NJ, NY, OR, PA, SC, WA, WI1995 – 1997 Added GA, HI, MO, TN, UT1998 – 2000 Added KY, ME, NC, TX, VA, WV2001 Added MI, MN, NE, RI, VT2002 Added NV, OH, SD(AZ data not available)2003 Added AZ, IN, NH(ME data not available)2004 Added AR(PA data not available)2005 Added OK(VA data not available)2006 Added ME, VA2007 Added WY2008 Added LA, PAFor the first NHQR, we calculated QI rates using two methods to test this hypothesis, first withdata from the full set of States in HCUP in 2000 and second with data from the set of States inHCUP in all three years (1994, 1997, and 2000), where that subset of States was re-weighted toobtain national estimates. For most QIs, the results differed very little.Non-Resident Discharges in State-level Estimates: HCUP databases include dischargesfrom all hospitals in a State, and may include non-residents, including foreign patients, whichcan bias the results for QIs using area-based denominators (State populations). We had noway to adjust the HCUP data to consistently exclude the non-resident discharges and includedischarges for residents hospitalized in other states. Therefore, non-resident discharges wereretained in the SID databases for the NHQR and NHDR analyses. Based on an analysisperformed with the 2007 SID, the percent of non-resident discharges is between 1% and 13%within a state. Most states were below 10%, but five states (NH, SD, TN, VT, WV) were above10%.Variation Among State QI Rates: Variation in State rates can be caused by many factors,including differences in practice patterns, underlying disease prevalence, health behaviors,access to health insurance, income levels of the population, demographics, spending on healthservices, supply of health care resources, coding conventions, and so on. To understand someof the variation in State rates, we analyzed the 2001 State rates in relation to these types offactors. Appendix G shows for each Prevention Quality Indicator (PQI) included in the NHQR,the analyses performed and the result in terms of whether the factors (with each testedseparately because of the limited number of observations) were positively, negatively, or notsignificantly related to the QIs.Methods for HCUP Data in 2011 NHQR and NHDR 8December 9, 2011

In a subsequent analysis, we investigated sources of variation in Patient Safety Indicator (PSI)rates across States using 2004 data. Appendix H contains the executive summary from thereport, Patient Safety in Hospitals in 2004: Toward Understanding Variation Across States. Theanalysis concluded there were few state factors (such as state policy, hospital characteristics,coding practices, and socio-demographics) with strong patterns of association to state-levelvariation in the nine PSI rates studied. The strongest result occurred with coding practices ―the number of diagnosis fields coded. Only one in five correlations between the PSIs and statefactors were statistically significant, although there is generally no pattern.These analyses are intended to help readers understand some of the external factors that maybe driving some of the State differences in PQI and PSI rates.Methods for HCUP Data in 2011 NHQR and NHDR 9December 9, 2011

TABLES

QI No.IQI 22IQI 23IQI 24IQI 25IQI 26IQI 27IQI 28IQI 29IQI 30DescriptionVaginal birth after cesarean (VBAC) per 1,000 women with previous cesarean deliveries (excludingpatients with abnormal presentation, preterm delivery, fetal death, multiple gestation diagnosis codesor breech procedure codes)Laparoscopic cholecystectomies per 1,000 cholecystectomy procedures (excluding complicated casesand obstetric and neonatal admissions), age 18 years and olderIncidental appendectomies per 1,000 patients with abdominal or pelvic surgery (excluding admissionsfor cancer of the appendix, admissions with a colectomy or pelvic evisceration, obstetric and neonataladmissions), age 65 years and olderBilateral cardiac catheterizations per 1,000 heart catheterizations for coronary artery disease(excluding valid indications for right-side catheterization and excluding obstetric and neonataladmissions)Coronary artery bypass grafts (excluding obstetric and neonatal admissions) per 100,000 population,age 40 years and olderPercutaneous transluminal coronary angioplasties (excluding obstetric and neonatal admissions) per100,000 population, age 40 years and olderHysterectomies (excluding obstetric and neonatal conditions, genital cancer, and pelvic or lowerabdominaltrauma) per 100,000 female population, age 18 years and olderLaminectomies or spinal fusions (excluding obstetric and neonatal conditions) per 100,000 population,age 18 years and olderDeaths per 1,000 adult admissions age 40 and older with percutaneous transluminal coronaryangioplasties (PTCA) (excluding obstetric and neonatal admissions and transfers to another hospital)IQI 31 Deaths per 1,000 admissions age 18 and older with carotid endarterectomies (CEA) (excludingobstetric and neonatal admissions and transfers to another hospital)IQI 32 Deaths per 1,000 admissions with acute myocardial infarction (AMI) as principal diagnosis (excludingtransfers from another hospital or to another hospital), age 18 years and olderIQI 33 First-time Cesarean deliveries per 1,000 deliveries (excluding patients with abnormal presentation,preterm delivery, fetal death, multiple gestation diagnosis codes, breech procedure codes, or aprevious Cesarean delivery diagnosis in any diagnosis field)IQI 34 Vaginal birth after cesarean (VBAC) per 1,000 women with previous cesarean deliveries with noexclusionsPatient Safety Indicators 6PSI 2 Deaths per 1,000 admissions with expected low-mortality* (excluding trauma, immunocompromised,and cancer patients), age 18 years or older or obstetric admissions* DRGs with a NIS 1997 benchmark of less than 0.5% mortality, excluding trauma,immunocompromised, and cancer patientsPSI 3 Pressure ulcers – Stage III or IV – per 1,000 discharges of length 5 or more days (excluding transfers;patients admitted from long-term-care facilities; patients with diseases of the skin, subcutaneoustissue, and breast; admissions for hemiplegia, paraplagia, quadriplagia, spina bifida, or anoxic braindamage; admissions in which debridement or pedicle graft is the only operating room procedure; andobstetrical admissions*), age 18 years or older* Also excludes admissions specifically for pressure ulcers, such as cases from earlier admissions orfrom other hospitals.PSI 4 Deaths per 1,000 elective-surgery admissions having developed specified complications of care*during hospitalizations of length 2 or fewer days (excluding patients transferred in or out, patientsadmitted from long-term-care facilities, and admissions specifically for specified complications of care),age 18 years to 89 years* Complications of care include acute renal failure, pneumonia, pulmonary embolism, deep veinthrombosis, sepsis, shock, cardiac arrest, gastroentestinal hemorrhage, and acute ulcer6 Indicators PSI 1 and 20 are not assigned by the PSI software, version 4. Incidence measures PSI 5(foreign body) and PSI 16 (transfusion reaction) are not calculated.Methods for HCUP Data in 2011 NHQR and NHDR 14December 9, 2011

QI No.PSI 6PSI 7PSI 8PSI 9PSI 10PSI 11PSI 12PSI 13PSI 14PSI 15PSI 17PSI 18DescriptionIatrogenic pneumothorax per 1,000 discharges (excluding obstetrical admissions and patients withchest trauma, pleural effusion, thoracic surgery, lung or pleural biopsy, diaphragmatic surgery repair,or cardiac surgery*), age 18 years or older* Also excludes admissions specifically for iatrogenic pneumothorax, such as cases from earlieradmissions or from other hospitals.Central venous catheter-related bloodstream infection per 1,000 medical and surgical discharges oflength 2 or more days (excluding immunocompromised and cancer patients, and admissionsspecifically for such infections*), age 18 years or older or obstetric admissions* Also excludes admissions specifically for such infections, such as cases from earlier admissions,from other hospitals, or from other settings.Postoperative hip fracture for adults per 1,000 surgical patients age 18 years and older who were notsusceptible to falling* (excluding obstetrical admissions)* That is, excluding patients admitted for seizures, syncope, stroke, coma, cardiac arrest, poisoning,trauma, delirium and other psychoses, anoxic brain injury; patients with metastatic cancer, lymphoidmalignancy, bone malignancy, and self-inflicted injury; admissions for diseases and disorders of themusculoskeletal system and connective tissue; and admissions in which hip fracture repair is the onlyoperating room procedure.Postoperative hemorrhage or hematoma with surgical drainage or evacuation, not verifiable asfollowing surgery*, per 1,000 surgical discharges (excluding obstetrical admissions), age 18 years orolder* Postoperative hemorrhage or hematoma is not verifiable as following surgery because information onday of procedure is not available for all discharges. Also, excludes admissions specifically for suchproblems, such as cases from earlier admissions, from other hospitals, or from other settings.Postoperative physiologic and metabolic derangements per 1,000 elective surgical discharges(excluding some serious disease* and obstetric admissions), age 18 years and older* That is, excluding patients with diabetic coma and patients with renal failure who also werediagnosed with AMI, cardiac arrhythmia, cardiac arrest, shock, hemorrhage, or gastrointestinalhemorrhage.Postoperative respiratory failure per 1,000 elective surgical discharges with an operating roomprocedure (excluding patients with respiratory disease, circulatory disease, neuromuscular disorders;obstetric conditions; admissions in which the tracheostomy is the only operating room procedure; andadmissions for craniofacial anomalies with laryngeal or pharyngeal surgery, or a procedure on face ),age 18 years and olderPostoperative pulmonary embolus (PE) or deep vein thrombosis (DVT) per 1,000 surgical discharges(excluding patients admitted for DVT, obstetrics, and interruption of vena cava before or aftersurgery*), age 18 years or older* Also excludes admissions specifically for such thromboembuli, such as cases from earlieradmissions, from other hospitals, or from other settings.Postoperative sepsis per 1,000 elective-surgery discharges with an operating room procedure oflength 4 or more days (excluding patients admitted for infection; patients with cancer orimmunocompromised states, obstetric conditions, and admissions specifically for sepsis), age 18 yearsor olderReclosure of postoperative abdominal wound dehiscence per 1,000 abdominopelvic-surgerydischarges of length 2 or more days (excluding immunocompromised patients, and obstetricconditions*), age 18 years or older* Also excludes admissions specifically for such wound dehiscence, such as cases from earlieradmissions or from other hospitals.Accidental puncture or laceration during procedures per 1,000 discharges (excluding obstetricadmissions and admissions involving spinal surgery*), age 18 years or older* Also excludes admissions specifically for such problems, such as cases from earlier admissions orfrom other hospitals.Birth trauma - injury to neonate per 1,000 live births (excluding preterm and osteogenesis imperfectabirths)Obstetric trauma with 3rd or 4th degree lacerations per 1,000 instrument-assisted vaginal deliveriesMethods for HCUP Data in 2011 NHQR and NHDR 15December 9, 2011

QI No.PSI 19PSI 21PSI 22PSI 23PSI 24PSI 25PSI 26PSI 27DescriptionObstetric trauma with 3rd or 4th degree lacerations per 1,000 vaginal deliveries without instrumentassistanceForeign body accidentally left in during procedure* per 100,000 population, age 18 years or older orobstetric admissions* Includes admissions specifically for treatment of foreign body left, such as cases from earlieradmissions or from other hospitals.Iatrogenic pneumothorax cases* per 100,000 population (excluding obstetrical admissions, andpatients with chest trauma, pleural effusion, thoracic surgery, lung or pleural biopsy, diaphragmaticsurgery repair, or cardiac surgery), age 18 years or older* Includes admissions specifically for iatrogenic pneumothorax, such as cases from earlier admissionsor from other hospitals. Also, includes barotrauma (including acute respiratory distress syndrome) andcentral line placement.Central venous catheter-related bloodstream infections* per 100,000 population (excludingimmunocompromised or cancer patients), age 18 years or older or obstetric admissions* Includes admissions specifically for such infections, such as cases from earlier admissions, fromother hospitals, or from other settings.Reclosure of postoperative abdominal wound dehiscence* (excluding immunocompromised andobstetric patients) per 100,000 population, age 18 years or older* Includes admissions specifically for such wound dehiscence, such as cases from earlier admissionsor from other hospitals.Accidental puncture or laceration during procedures* per 100,000 population (excluding obstetricadmissions and admissions involving spinal surgery), age 18 years or older* Includes admissions specifically for such problems, such as cases from earlier admissions or fromother hospitals.Transfusion reactions* per 100,000 population (excluding neonates), age 18 years or older or obstetricadmissions* Includes admissions specifically for transfusion reactions, such as cases from earlier admissions orfrom other hospitals.Postoperative hemorrhage or hematoma with surgical drainage or evacuation, not verifiable asfollowing surgery* (excluding obstetrical admissions), per 100,000 population, age 18 years or older* Postoperative hemorrhage or hematoma is not verifiable as following surgery because information onday of procedure is not available for all discharges. Also, includes admissions specifically for suchproblems, such as cases from earlier admissions or from other hospitals.Methods for HCUP Data in 2011 NHQR and NHDR 16December 9, 2011

Table 2. Sources of 2008 HCUP Data for the NHQR and the NHDRSponsored by the Agency for Healthcare Research and Quality (AHRQ), HCUP is a family ofdatabases, software tools, and products developed through the collaboration of State dataorganizations, hospital associations, private data organizations, and the Federal government.HCUP would not be possible without the contributions of the following data collection Partnersfrom across the United States:Data Sources for the HCUP Nationwide Inpatient Sample andState Inpatient DatabasesArizona Department of Health ServicesArkansas Department of HealthCalifornia Office of Statewide Health Planning and DevelopmentColorado Hospital AssociationConnecticut Hospital AssociationFlorida Agency for Health Care AdministrationGeorgia Hospital AssociationHawaii Health Information CorporationAlso included in thedisparities analysis filesIllinois Department of Public Health ---Indiana Hospital Association ---Iowa Hospital Association ---Kansas Hospital AssociationKentucky Cabinet for Health and Family ServicesLouisiana Department of Health and Hospitals ---Maine Health Data OrganizationMaryland Health Services Cost Review CommissionMassachusetts Division of Health Care Finance and PolicyMichigan Health & Hospital AssociationMinnesota Hospital Association ---Missouri Hospital Industry Data InstituteNebraska Hospital Association ---Nevada Department of Health and Human ServicesNew Hampshire Department of Health & Human ServicesNew Jersey Department of Health and Senior ServicesNew York State Department of HealthNorth Carolina Department of Health and Human Services ---Ohio Hospital Association ---Oklahoma State Department of HealthOregon Association of Hospitals and Health SystemsPennsylvania Health Care Cost Containment CouncilRhode Island Department of HealthYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesYesMethods for HCUP Data in 2011 NHQR and NHDR 17December 9, 2011

Data Sources for the HCUP Nationwide Inpatient Sample andState Inpatient DatabasesSouth Carolina State Budget & Control BoardAlso included in thedisparities analysis filesSouth Dakota Association of Healthcare Organizations ---Tennessee Hospital AssociationTexas Department of State Health ServicesUtah Department of HealthVermont Association of Hospitals and Health SystemsVirginia Health InformationWashington State Department of Health ---West Virginia Health Care Authority ---Wisconsin Department of Health ServicesWyoming Hospital AssociationYesYesYesYesYesYesYesYesMethods for HCUP Data in 2011 NHQR and NHDR 18December 9, 2011

Table 3. Age Groupings for Risk AdjustmentThis table shows the 18 categories of patient age, in five-year increments, that are used for riskadjustment. The 36 age-gender categories for risk adjustment are constructed from the 18 agecategories split into male-female gender.Age Groups0-45-910-1415-1718-2425-2930-3435-3940-4445-4950-5455-5960-6465-6970-7475-7980-8485 or olderMethods for HCUP Data in 2011 NHQR and NHDR 19December 9, 2011

Table 4. Use of Secondary Procedure Days in AHRQ Quality Indicators, Version 4.1Eight PSIs and four PDIs used information on the timing of procedures (PRDAY) to excludepatients:PSI 3 – Pressure UlcerPSI 8 – Post-operative hip fracturesPSI 9 – Post-operative hemorrhage or hematomaPSI 10 – Post-operative physiologic/metabolic derangementsPSI 11 – Post-operative respiratory failurePSI 12 – Post-operative pulmonary embolism or deep vein thrombosisPSI 14 – Post-operative abdominal wound dehiscencePSI 27 – Post-operative hemorrhage or hematoma (area based)PDI 2 – Pediatric: Pressure ulcerPDI 8 – Pediatric: Post-operative hemorrhage or hematomaPDI 9 – Pediatric: Post-operative respiratory failurePDI 11 – Pediatric: Post-operative wound dehiscenceMethods for HCUP Data in 2011 NHQR and NHDR 20December 9, 2011

Table 5. Use of Present on Admission in AHRQ Quality Indicators, Version 4.1Fourteen PSIs and 16 PDIs used information on whether a condition was present onadmission (POA) to exclude patients:PSI 3 Pressure UlcerPSI 6 Iatrogenix PneumothoraxPSI 7 Central Venous Catheter-Related Bloodstream InfectionPSI 8 Postoperative Hip FracturePSI 9 Postoperative Hemorrhage or HematomaPSI 10 Postoperative Physiologic and Metabolic DerangementsPSI 11 Postoperative Respiratory FailurePSI 12 Postoperative Pulmonary Embolism or Deep Vein ThrombosisPSI 13 Postoperative SepsisPSI 14 Postoperative Wound Dehiscence (Provider-based)PSI 15 Accidental Puncture or Laceration (Provider-based)PDI 1 Pediatric: Accidental Puncture or LacerationPDI 2 Pediatric: Pressure UlcerPDI 5 Pediatric: Iatrogenic PneumothoraxPDI 8 Pediatric: Postoperative Hemorrhage or HematomaPDI 9 Pediatric: Postoperative Respiratory FailurePDI 10 Pediatric: Postoperative SepsisPDI 11 Pediatric: Reclosure of postoperative abdominal wound dehiscencePDI 12 Pediatric: Central Venous Catheter-Related Bloodstream InfectionMethods for HCUP Data in 2011 NHQR and NHDR 21December 9, 2011

Table 6. Number of Diagnosis and Procedure Fields by State, 2008StateMaximum Number ofDiagnosesMaximum Number ofProceduresArkansas 18 8Arizona 25 12California 25 21Colorado 15 15Connecticut 30 30Florida 31 31Georgia 30 30Hawaii 20 20Illinois 9 6Indiana 15 15Iowa 66 37Kansas 30 25Kentucky 25 25Louisiana 9 7Maine 10 6Maryland 30 15Massachusetts 15 15Michigan 30 30Minnesota 25 25Missouri 30 25Nebraska 9 6Nevada 33 12New Hampshire 10 6New Jersey 25 24New York 15 15North Carolina 24 8Ohio 15 9Oklahoma 16 16Oregon 25 25Pennsylvania 9 6Rhode Island 25 25South Carolina 15 13South Dakota 61 46Tennessee 18 6Texas 25 15Utah 9 6Vermont 20 20Virginia 18 6Washington 25 25West Virginia 18 6Wisconsin 30 30Wyoming 30 25Methods for HCUP Data in 2011 NHQR and NHDR 22December 9, 2011

REFERENCESAgency for Healthcare Research and Quality. Prevention Quality Indicators: SoftwareDocumentation, Version 4.1 SAS. Rockville, MD: Agency for Healthcare Research and Quality,2009.Agency for Healthcare Research and Quality. Inpatient Quality Indicators: SoftwareDocumentation, Version 4.1 SAS. Rockville, MD: Agency for Healthcare Research and Quality,2009.Agency for Healthcare Research and Quality. Patient Safety Indicators: SoftwareDocumentation,Version 4.1 SAS. Rockville, MD: Agency for Healthcare Research and Quality,2009.Agency for Healthcare Research and Quality. Pediatric Quality Indicators: SoftwareDocumentation,Version 4.1 SAS. Rockville, MD: Agency for Healthcare Research and Quality,2009.Claritas, Inc. The Claritas Demographic Update Methodology, April 2008.Coffey R, Barrett M, Houchens R, Moy E, Ho K, Andrews R, Moles E. Methods Applying AHRQQuality Indicators to Healthcare Cost and Utilization Project (HCUP) Data for the Seventh(2009) National Healthcare Disparities Report. HCUP Methods Series Report # 2009-01. OnlineAugust 17, 2009. U.S. Agency for Healthcare Research and Quality. Available: http://www.hcupus.ahrq.gov/reports/methods.jsp.Fleiss JL. Statistical Methods for Rates and Proportions. New York: Wiley, 1973.Houchens R, Elixhauser A. Final Report on Calculating Nationwide Inpatient Sample (NIS)Variances, 2001. HCUP Methods Series Report #2003-2. Online June 2005 (revised June 6,2005). U.S. Agency for Healthcare Research and Quality. Available: http://www.hcupus.ahrq.gov/reports/methods.jsp.IOM (Institute of Medicine). Future Directions for the National Healthcare Quality and DisparitiesReports. Washington, DC: The National Academies Press, 2010.Raetzman S, Stranges E, Coffey RM, Barrett ML, Andrews R, Moy E, Brady J. Patient Safety inHospitals in 2004 : Toward Understanding Variation Across States. HCUP Methods SeriesReport #2008-2. Online March 2008. U.S. Agency for Healthcare Research and Quality.Available: http://www.hcup-us.ahrq.gov/reports/methods.jsp.Methods for HCUP Data in 2011 NHQR and NHDR 24December 9, 2011

APPENDICES

APPENDIX A:DEVELOPMENT OF THE DISPARITIES ANALYSIS FILEFOR NATIONAL QI ESTIMATES BY RACE/ETHNICITYRace and ethnicity measures can be problematic in hospital discharge databases becausemany hospitals do not code race and ethnicity completely. Because race/ethnicity is a pivotalmeasure for the NHDR, we explored the reporting practices in the 42 States that participate in2008 HCUP SID. Six States did not provide information on patient race to HCUP. Five Statesdid not report Hispanic ethnicity. The remaining 31 States were used for the creation of thedisparities analysis files (See Table 2 in the main body of the report for the list of States).The following table demonstrates the representation by U.S. Census region of these 31 States.CensusRegionNumber of States usedfor the disparitiesanalysis fileNumber of States inthe regionPercent of States in theregion included in thedisparities analysis fileNortheast 9 9 100%Midwest 4 12 33%South 10 16 63%West 7 13 54%Total 31 50 62%The table below compares aggregated totals of various measures for the 31 States as a percentof the national measure. In 2008, the 31 States accounted for 69 percent of U.S. hospitaldischarges (based on the American Hospital Association’s Annual Survey). They accounted forabout 71 percent of White and African Americans in the nation and 87 percent of Asian/PacificIslanders and Hispanics (based on 2008 Claritas data).MeasureTotal of 31 HCUP States with race/ethnicityas a percent of national totalHospital discharges 69%Total resident population 74%*Population by race/ethnicity:White 71%*African American 71%*Asian/Pacific Islander 86%*Hispanic 87%*Population by age:Population under age 18 74%*Population age 18-64 74%*Population over age 64 74%*Population by income:Population with income under the poverty level 71%***Calculated using 2008 Claritas data and 1977 OMB Directive 15 race definitions (e.g. no option forselecting “two or more races”).**Calculated using Kaiser Family Foundation, statehealthfacts.org. Data Source: Urban Institute andKaiser Commission on Medicaid and the Uninsured estimates based on the Census Bureau's March2009and 2010 Current Population Survey (CPS: Annual Social and Economic Supplements),accessed on September 28, 2011.Methods for HCUP Data in 2011 NHQR and NHDR A-1December 9, 2011

HCUP Race DataHCUP coding includes race and ethnicity in one data element (RACE). Because of variability inthe collection of race and ethnicity information in the State data provided to HCUP, HCUPmaintains a uniform set of categories based on race definitions used in the 1977 Office ofManagement and Budget (OMB) Directive 15 (separate categories for Hispanic and five Non-Hispanic racial groups – White, Black, Asian or Pacific Islander, Native American, and Other).When a State and its hospitals collect Hispanic ethnicity separately from race, HCUP assignsthe data to the combined race/ethnicity categorization and uses Hispanic ethnicity to overrideany other race category to create uniform coding across states. Because of limited reporting ofNative American (American Indian/Alaska Native) in the HCUP data, counts for Other andAmerican Indian/Alaska Native are combined into “Other” races for the NHDR analyses.Preparing the Disparities Analysis FileThe sampling and weighting strategy used for the disparities analysis file for national estimatesby race/ethnicity is similar to the method used to create the HCUP NIS, except that thedisparities analysis file draws its sample from 31 of the 42 States included in the 2008 NIS andis a 40-percent sample of community hospitals rather than a 20-percent sample as in the NIS. First, community hospitals from the 31 States were sampled to approximate a 40-percent stratified sample of U.S. community hospitals. The sampling strata were definedbased on five hospital characteristics: geographic region, hospital control (i.e., public,private not-for-profit, and proprietary), urbanized location, teaching status, and bed size.Hospitals were excluded from the sampling frame if the coding of patient race wassuspect (i.e., more than 30% of the discharges in the hospital had the race reported as“other”; more than 50% of the discharges had no information on the race of the patient;all of the discharges in the hospital had race coded as white, other, or missing; or 100%of the discharges had race coded as white and the hospital had more than 50 beds).For discharges missing race, a “hot deck” imputation method (which draws donors fromstrata of similar patients within the same hospital) is used to assign values whilepreserving the variance within the data.Once the 40-percent sample was drawn, discharge-level weights were developed toproduce national-level estimates when applied to the disparities analysis file.The final disparities analysis file included about 15 million hospital discharges from close to2,000 hospitals.The 2010 NHDR also reports information derived from the 2001, 2004, 2005, and 2007disparities analysis files for comparison. These additional data files were developed using theyear-specific SID and the same approach described above. QI statistics for the back yearswere re-run using the modified version 4.1 software to that the same version of the QI softwareis used for all years in a given NHQR release.Evaluating the Disparities Analysis FileAfter creating the 2008 disparities analysis file using the above steps, we evaluated thereliability of national estimates produced with these data by comparing its composition to the2008 NIS. The tables below contain the distribution of discharges in both files by keyMethods for HCUP Data in 2011 NHQR and NHDR A-2December 9, 2011

demographic and clinical data elements. Based on these analyses, the 2008 disparitiesanalysis file appears to provide reliable national estimates when compared with the NIS.Weighted FrequenciesStratum used to sample hospitals2008 DisparitiesAnalysis File2008 NISNHDR_STRATUM Frequency Percent Frequency Percent1: Northeast 7,747,709 12.3% 7,747,709 12.3%2: Midwest 9,208,364 20.9% 9,208,364 20.9%3: South 15,181,340 34.7% 15,181,340 34.7%4: West 6,549,630 16.0% 6,549,630 16.0%Age in years at admission2008 DisparitiesAnalysis File2008 NISAGE Frequency Percent Frequency Percent.: Missing 4,980 0.0 45,400 0.1.A: Invalid 176 0.0 205 0.0.C: Inconsistent 6,586 0.0 4,096 0.00-17 6,686,033 16.8 6,349,289 15.918-44 10,156,359 25.5 10,025,574 25.145-64 9,433,838 23.7 9,504,242 23.865+ 13,597,149 34.1 13,956,313 35.0Indicator of sex2008 DisparitiesAnalysis File2008 NISFEMALE Frequency Percent Frequency Percent.: Missing 2,410 0.0 108,451 0.3.A: Invalid 148 0.0 185 0.0.C: Inconsistent 1,723 0.0 2,612 0.00: Male 16,550,967 41.5 16,498,753 41.41: Female 23,329,871 58.5 23,275,119 58.4Methods for HCUP Data in 2011 NHQR and NHDR A-3December 9, 2011

Primary expected payer2008 DisparitiesAnalysis File2008 NISPAY1 Frequency Percent Frequency Percent.: Missing 57,515 0.1 59,363 0.1.A: Invalid 7,090 0.0 8,240 0.01: Medicare 14,837,833 37.2 14,917,422 37.42: Medicaid 8,030,521 20.1 7,354,724 18.43: Private Insurance 13,506,103 33.9 14,108,326 35.44: Self-pay 1,985,225 5.0 1,912,058 4.85: No Charge 187,139 0.5 214,220 0.56: Other 1,273,695 3.2 1,310,766 3.3Patient race/ethnicity 72008 DisparitiesAnalysis File2008 NISRACE Frequency Percent Frequency Percent.: Missing 697,511 1.7 8,082,120 20.3.A: Invalid 3,694 0.0 9,144 0.01: White 26,428,543 66.3 21,640,795 54.32: Black 5,725,413 14.4 4,146,702 10.43: Hispanic 4,768,632 12.0 3,707,280 9.34: Asian/Pacific Islander 1,000,988 2.5 908,217 2.35: Native American 277,057 0.7 216,578 0.56: Other 983,283 2.5 1,174,285 2.9Location of patient residence2008 DisparitiesAnalysis File2008 NISPL_NCHS Frequency Percent Frequency Percent.: Missing 0 0 1,242,853 3.11: Large central metro 12,027,765 30.2 12,072,988 30.32: Large fringe metro 10,099,290 25.3 8,968,306 22.53: Medium metro 6,920,105 17.4 6,986,756 17.54: Small metro 3,620,426 9.1 3,325,294 8.35: Micropolitan (nonmetro) 4,088,133 10.2 4,455,368 11.26: Noncore (nonmetro) 3,129,401 7.8 2,833,555 7.17 Differences in race distribution are attributable to high rates of missing race on the NIS (20%). The2008 disparities analysis file uses a modified race variable with missing or invalid values imputed andNative American and Other combined into one racial group.Methods for HCUP Data in 2011 NHQR and NHDR A-4December 9, 2011

Top 24 DRGs(Combination of Top 24 DRGs for Disparities and NIS file)2008 DisparitiesAnalysis File2008 NISDRG, Version 25Frequency Percent Frequency Percent795: NORMAL NEWBORN 3,146,389 7.9 3,069,455 7.7775: VAGINAL DELIVERY W/O COMPLICATING 2,449,302 6.1 2,387,788 6.0DIAGNOSES885: PSYCHOSES 1,097,381 2.8 1,079,025 2.7766: CESAREAN SECTION W/O CC/MCC 946,931 2.4 904,258 2.3392: ESOPHAGITIS, GASTROENT & MISC DIGEST 841,636 2.1 865,121 2.2DISORDERS W/O MCC470: MAJOR JOINT REPLACEMENT OR834,007 2.1 912,097 2.3REATTACHMENT OF LOWER EXTREMITYW/O MCC794: NEONATE W OTHER SIGNIFICANT709,055 1.8 701,674 1.8PROBLEMS313: CHEST PAIN 583,335 1.5 605,633 1.5603: CELLULITIS W/O MCC 490,121 1.2 488,359 1.2765: CESAREAN SECTION W CC/MCC 480,637 1.2 472,330 1.2871: SEPTICEMIA W/O MV 96+ HOURS W MCC 477,282 1.2 472,912 1.2194: SIMPLE PNEUMONIA & PLEURISY W CC 472,053 1.2 470,397 1.2641: NUTRITIONAL & MISC METABOLIC446,247 1.1 451,807 1.1DISORDERS W/O MCC743: UTERINE & ADNEXA PROC FOR NON-440,584 1.1 429,963 1.1MALIGNANCY W/O CC/MCC690: KIDNEY & URINARY TRACT INFECTIONS 418,595 1.0 432,837 1.1W/O MCC247: PERC CARDIOVASC PROC W DRUG-364,593 0.9 389,818 1.0ELUTING STENT W/O MCC287: CIRCULATORY DISORDERS EXCEPT AMI, W 349,822 0.9 358,593 0.9CARD CATH W/O MCC291: HEART FAILURE & SHOCK W MCC 345,909 0.9 346,599 0.9292: HEART FAILURE & SHOCK W CC 341,687 0.9 342,736 0.9774: VAGINAL DELIVERY W COMPLICATING339,986 0.9 342,341 0.9DIAGNOSES195: SIMPLE PNEUMONIA & PLEURISY W/O337,220 0.8 333,989 0.8CC/MCC945: REHABILITATION W CC/MCC 336,724 0.8 303,770 0.8203: BRONCHITIS & ASTHMA W/O CC/MCC 328,736 0.8 310,221 0.8192: CHRONIC OBSTRUCTIVE PULMONARYDISEASE W/O CC/MCC328,250 0.8 334,821 0.8Methods for HCUP Data in 2011 NHQR and NHDR A-5December 9, 2011

Median income of Patient’s ZIP Code2008 DisparitiesAnalysis File2008 NISZIPINC_QRTL Frequency Percent Frequency Percent.: Missing 0 0 819,827 30.31: First Quartile (lowest income) 12,065,933 30.3 11,007,621 26.42: Second Quartile 10,534,029 26.4 10,745,699 22.53: Third Quartile 8,989,251 22.5 8,934,411 20.84: Fourth Quartile (highest income) 8,295,907 20.8 8,375,158 30.3A: Invalid 0 0 2,404 0.0Weighted MeansVariable / LabelLOS: Length ofstay (cleaned)NDX: Number ofdiagnoses onthis recordNPR: Number ofprocedures onthis recordTOTCHG: Totalcharges (cleaned)2008 Disparities Analysis File 2008 NISMinimum Maximum Mean Minimum Maximum Mean0 365 4.7 0 364 4.60 33 7.6 0 66 7.30 31 1.6 0 32 1.6$100 $1,499,860 $29,272.76 $100 $1,499,552 $29,047.95Methods for HCUP Data in 2011 NHQR and NHDR A-6December 9, 2011

APPENDIX B:DEVELOPMENT OF THE DISPARITIES ANALYSIS FILESFOR STATE-LEVEL QI ESTIMATES BY RACE/ETHNICITYData from the 2008 SID were used to create individual state disparities analysis files that weredesigned to provide State-level QI estimates by race/ethnicity. The starting point for State-leveldisparities analysis files were the SID prepared for the other reporting in the NHQR, asdescribed in the HCUP Databases section of this report. These files were limited to community,non-rehabilitation hospitals with missing characteristics imputed. Disparities analysis files werecreated for the 31 HCUP States that report race/ethnicity of discharges (see Table 2 in the mainbody of the report for a list of the States).The following steps were taken to further prepare the State-level files for reporting byrace/ethnicity:1. Selection of Hospitals. We first selected hospitals whose original coding of patient raceethnicitywas not “suspect.” Hospitals were removed from the State-level disparities analysisfiles if the quality of the race-ethnicity reporting was suspect, using the same four criteria forexclusion of hospitals with suspect race coding that were applied when creating the nationaldisparities analysis file (see Appendix A for details).In 24 of the 31 States with race/ethnicity data, at least one hospital was eliminated due tosuspect race coding. Seven had no hospitals with suspect race coding. Overall, 5.1percent of hospitals and 2.8 percent of discharges were excluded. The table belowindicates the reason for excluding hospitals and their associated discharges from the Stateleveldisparities analysis files. Except in a few cases, hospitals in a State were most oftenexcluded because substantial shares of discharges were coded as “other” or “missing” race.MeasureTotal numberof hospitalsexcludedTotal numberof dischargesexcludedExclusions from State-level Disparities Analysis Files for Race/EthnicityExcludedfor anyreasonPercentof Total>30%dischargesare "other"race>50%dischargesare"missing"raceAlldischargesare white,other ormissingAlldischargesare whiteandhospitalhas >50beds161 5.1% 72 66 22 1778,053 2.8% 321,192 447,668 9,174 192. Impute for Missing Race/Ethnicity. Because the area-level measures selected for this reportuse total state population in the denominator, minimizing the loss of discharges from thenumerator for the QI calculation is critical to producing unbiased QI rates. For missing race,we used a “hot deck” imputation method (which draws donors from strata of similar patientswithin the same hospital) to assign values while preserving the variance within the data.Methods for HCUP Data in 2011 NHQR and NHDR B-1December 9, 2011

Typically, most States have no more than seven percent (7%) of discharges starting out withmissing race values.3. Weighting of Selected Hospitals. We calculated discharge-level weights to account forhospitals excluded because of suspect race coding, community hospitals not reported in theSID, and missing quarters of data. We weighted to the State’s universe of hospitals in theAmerican Hospital Association (AHA) Annual Survey Database based on hospitalcharacteristics.There may be differences in race and ethnicity coding among States that affect the estimates.For example, some States include Hispanic ethnicity as one of the racial categories, and othersrecord Hispanic ethnicity separately from race. At the hospital-level, policies vary on methodsfor collecting such data. Some hospitals ask the patient to identify their race and ethnicity, andothers determine it from observation. The effect of these and other unmeasured differences incoding of race and ethnicity across the States and hospitals cannot be assessed.Methods for HCUP Data in 2011 NHQR and NHDR B-2December 9, 2011

APPENDIX C:STATISTICAL METHODSThis appendix explains the statistical methods and gives formulas for the calculations ofstandard errors and hypothesis tests. These statistics are derived from multiple databases: theNIS, the SID, and Claritas (a vendor that compiles and adds value to Bureau of Census data).For NIS estimates and the disparities analysis file, the standard errors are calculated asdescribed in the HCUP report entitled Calculating Nationwide Inpatient Sample (NIS) Variances(Houchens, et al., 2005). We will refer to this report simply as the NIS Variance Reportthroughout this appendix. This method takes into account the cluster and stratification aspectsof the NIS sample design when calculating these statistics using the SAS procedure PROCSURVEYMEANS. For the SID we used the same procedure omitting the cluster andstratification features. For population counts based on Claritas data, there is no sampling error.Even though the NIS and the disparities analysis file contain discharges from a finite sample ofhospitals and most of the SID databases contain nearly all discharges from nearly all hospitalsin the state, we treat the samples as though they were drawn from an infinite population. We donot employ finite population correction factors in estimating standard errors. We take thisapproach because we view the outcomes as a result of myriad processes that go into treatmentdecisions rather than being the result of specific, fixed processes generating outcomes for aspecific population and a specific year. We consider the NIS and SID to be samples from a“super-population” for purposes of variance estimation. Further, we assume the counts (of QIevents) to be binomial.1. Area Population QIs using Claritas Population Dataa. Standard error estimates for discharge rates per 100,000 population using the 2008Claritas population data.The observed rate was calculated as follows:R 100,000 ni1w xNii 100,000 SN.(A.1)w i and x i, respectively, are the weight and variable of interest for patient i in the NIS or SID.To obtain the estimate of S and its standard error, SE S , we followed instructions in the NISVariance Report (modified for the SID, as explained above)The population count in the denominator is a constant. Consequently, the standard error ofthe rate R was calculated as:SE R =100,000 SE S / N.(A.2)Methods for HCUP Data in 2011 NHQR and NHDR C-1December 9, 2011

. Standard error estimates for age/sex adjusted inpatient rates per 100,000 populationusing the 2008 Claritas population data.We adjusted rates for age and sex using the method of direct standardization (Fleiss, 1973).We estimated the observed rates for each of 36 age/sex categories (described in Table 3 inthis methods report, Age Groupings for Risk Adjustment). We then calculated the weightedaverage of those 36 rates using weights proportional to the percentage of a standardpopulation in each cell. Therefore, the adjusted rate represents the rate that would beexpected for the observed study population if it had the same age and sex distribution as thestandard population.For the standard population we used the age and sex distribution of the U.S. as a wholeaccording to the year 2000. In theory, differences among adjusted rates were notattributable to differences in the age and sex distributions among the comparison groupsbecause the rates were all calculated with a common age and sex distribution.The adjusted rate was calculated as follows (and subsequently multiplied by 100,000):A 36g 1Ng , std36g 1n(g )i1 g , obsNwg,stdg,iNxg,i36g 1n(g )i1NNNg,stdg , obsstdwg,ixg , i36g 1n(g )i1Nwstd*g , ixg , iS *Ng = index for the 36 age/sex cells.N g,std = Standard population for cell g (year 2000 total US population in cell g).N g,obs = Observed population for cell g (year 2008 subpopulation in cell g, e.g., females,state of California, etc.).n(g) = Number in the sample for cell g.x g,i = Observed quality indicator for observation i in cell g (e.g., 0 or 1 indicator).w g,i = NIS or SID discharge weight for observation i in cell g.The estimates for the numerator, S*, and its standard error, SE S* , were calculated in similarfashion to the unadjusted estimates for the numerator S in formula A.1. The only differencewas that the weight for patient i in cell g was redefined as:* g,stdwg, i wg,iNg,obsstd.(A.3)N (A.4)Following instructions in the NIS Variance Report (modified for the SID, as explainedabove), we used PROC SURVEYMEANS to obtain the estimate of S*, the weighted sum inthe numerator using the revised weights, and the estimate SE S* , the standard error of theweighted sum S*. The denominator is a constant. Therefore, the standard error of theadjusted rate, A, was calculated asSE A =100,000 SE S* / N std .(A.5)Methods for HCUP Data in 2011 NHQR and NHDR C-2December 9, 2011

2. Provider-based QIs using Weighted Discharge Data (SID and NIS)a. Standard error estimates for inpatient rates per 1,000 discharges using dischargecounts in both the numerator and the denominator.We calculated the observed rate as follows:R 1,000 ni1ni1w xiwii 1,000 SN.(A.6)Following instructions in the HCUP NIS Variance Report (modified for the SID, as explainedabove), we used PROC SURVEYMEANS to obtain estimates of the weighted mean, S/N,and the standard error of the weighted mean, SE S/N . We multiplied this standard error by1,000.b. Standard error estimates for age/sex adjusted inpatient rates per 1,000 dischargesusing inpatient counts in both the numerator and the denominator.We used the 2000 NIS national estimates for the standard inpatient population age-sexdistribution. For each of the 36 age-sex categories, we estimated the number of U.S.inpatient discharges, ˆ , in category g. We calculated the directly adjusted rate:Ng , stdw36g,ixg,iˆ i1 Ng,std n(g )n(g )g 1 wg,i36 wg,ixi1A 1,0001,000 ˆ i1 36 Pg, std n(g )ˆg1Nwg1n(g )g,stdi1g,ig,i.(A.7)g = index for the 36 age/sex cells.N ˆ g , std= Standard inpatient population for cell g (Estimate of year 2000 total inpatientpopulation for cell g).n(g) = Number in the sample for cell g.x g,i = Observed quality indicator for observation i in cell g.w g,i = NIS or SID discharge weight for observation i in cell g.ˆ,Note that ˆNg stdPg, stdis the proportion of the standard inpatient population in cell g.Nˆ 36g1g,stdConsequently, the adjusted rate is a weighted average of the cell-specific rates with cellweights equal to ˆ . These cell weights are merely a convenient, reasonable standardPg , stdMethods for HCUP Data in 2011 NHQR and NHDR C-3December 9, 2011

inpatient population distribution for the direct standardization. Therefore, we treat these cellweights as constants in the variance calculations:n(g )n(g ) 36 wg,ixg,i36 wg,ixg,i( )1,000 ˆ i11,000 ˆ 2 i1SE A Var(A) Var, ,.Pgstd ( ) PgstdVarn g n(g ) g1g1 wg,i wg,i i1 i1(A.8)The variance of the ratio enclosed in parentheses was estimated separately for each cell gby squaring the SE calculated using the method of section 2.a:SE(A)1,000Rgn(g)wg,ii1n(g)i1wxg,ig,i36g12Pˆg, std SE(Rg)Following instructions in the HCUP NIS Variance Report (modified for the SID, as explainedabove), we used PROC SURVEYMEANS to obtain estimates of the weighted means, R g ,and their standard errors.3. Significance tests.2(A.9)Let R 1 and R 2 be either observed or adjusted rates calculated for comparison groups 1 and2, respectively. Let SE 1 and SE 2 be the corresponding standard errors for the two rates. Wecalculated the test statistic and (two-sided) p-value:R1 R2t 2 2SE1 SE2(A.10)p 2 * Prob( Z t )where Z is a standard normal variate.Note: the following functions calculate p in SAS and EXCEL:SAS: p = 2 * (1 - PROBNORM(ABS(t)));EXCEL: = 2*(1- NORMDIST(ABS(t),0,1,TRUE))Methods for HCUP Data in 2011 NHQR and NHDR C-4December 9, 2011

APPENDIX D:NHQR/NHDR SUMMARY MEASURES FORPATIENT SAFETY AND MORTALITY FOR SELECTEDPROCEDURES AND CONDITIONSTo examine national and state-level trends in inpatient mortality and patient safety events, riskadjustedrates for select AHRQ Inpatient Quality Indicators (IQIs) and Patient Safety Indicators(PSIs) were summarized. The three NHQR/NHDR summary measures include the following:1. Mortality for selected conditions based on select IQIs2. Mortality for selected procedures based on select IQIs3. Patient safety based on select PSIsThese summary measures were calculated as a weighted sum of risk-adjusted rates forindividual IQIs and PSIs. The weights used to calculate the NHQR/NHDR summary measureswere relatively consistent with AHRQ IQI and PSI Composites; however, the methodologyemployed to perform the calculations differed. The IQI and PSI composites were designed foruse with hospital-level rates, while the NHQR/NHDR report only national and state-levelstatistics.The NHQR/NHDR summary measure for mortality for selected conditions was based on six IQIsalso included in the similar IQI Composite.IQI CompositeWeightNHQR/NHDRSummary MeasureWeightIQI DescriptionIQIs Included in the NHQR/NHDR SummaryIQI15 Acute Myocardial Infarction 0.1433 0.1433IQI16 Congestive Heart Failure 0.2739 0.2739IQI17 Acute Stroke Adult Mortality Rate 0.1329 0.1329IQI18 Gastrointestinal Hemorrhage 0.1302 0.1302IQI19 Hip Fracture 0.0678 0.0678IQI20 Pneumonia 0.2519 0.2519The IQI composite weights were extracted from the SAS software, version 4.1. They are basedon pooled SID denominators (i.e., the relative frequency of the denominators of the componentindicators). This approach is known as “opportunity weighting,” because it gives equal weight toeach opportunity that the health care system has to do “the right thing,” which in this case is todischarge the patient alive from the hospital. The NHQR/NHDR summary measure weightswere the same as the weights in the similar IQI Composite.The NHQR/NHDR summary measure for mortality for selected procedures was based on fourIQIs instead of the eight IQIs included in the similar IQI Composite.Three IQIs were excluded because the procedures were not high-volume at the state level and,therefore, state-level risk-adjusted rates were often unavailable. The IQI for Hip Replacementhas a zero-weight in the IQI Composite because it is not endorsed by the National QualityForum. The IQI composite weights were extracted from the SAS software, version 4.1, andwere also based on pooled SID denominators. The IQI Composite weights were proportionallyreallocated into the NHQR/NHDR summary measure weights to account for the excluded IQIs.Methods for HCUP Data in 2011 NHQR and NHDR D-1December 9, 2011

IQI CompositeWeightNHQR/NHDRSummary MeasureWeightIQI DescriptionIQIs Included in the NHQR/NHDR SummaryIQI30 PTCA 0.5659 0.6275IQI12 CABG 0.2001 0.2219IQI13 Craniotomy 0.1031 0.1143IQI11 Abdominal Aortic Aneurysm Repair 0.0328 0.0364IQIs Excluded in the NHQR/NHDR Summary, but in the IQI CompositeIQI08 Esophageal Resection 0.0043 0.0000IQI09 Pancreatic Resection 0.0048 0.0000IQI14 Hip Replacement 0.0000 0.0000IQI31 Carotid Endarterectomy 0.0890 0.0000The NHQR/NHDR summary measure for patient safety was based on seven PSIs instead of theeleven PSIs included in the similar PSI Composite.NHQR/NHDRSummary MeasureWeightPSI DescriptionPSI CompositeWeightPSIs Included in the NHQR/NHDR SummaryPSI15 Accidental Puncture or Laceration 0.2982 0.3925Postoperative Pulmonary Embolism orPSI12 Deep Vein Thrombosis 0.2360 0.3106Central Venous Catheter-RelatedBloodstream Infections (2008 only) 0.1280 0.1685PSI07PSI06 Iatrogenic Pneumothorax 0.0457 0.0602PSI13 Postoperative Sepsis (2008 only) 0.0383 0.0504PSI14 Postoperative Wound Dehiscence 0.0124 0.0163PSI08 Postoperative Hip Fracture 0.0011 0.0014PSIs Excluded in the NHQR/NHDR Summary, but in the PSI CompositePSI03 Pressure Ulcer 0.2403 0.0000PSI09Postoperative Hemorrhage orHematoma 0.0000 0.0000PSI10Postoperative Physiologic andMetabolic Derangement 0.0000 0.0000PSI11 Postoperative Respiratory Failure 0.0000 0.0000One PSI Pressure Ulcer was excluded due to its dependence upon reporting whether thediagnosis is present on admission (POA) to the hospital. (This information is not uniformlyavailable across HCUP States). Three PSIs have zero weights in the PSI Composite becausethey are not endorsed by the National Quality Forum. The PSI composite weights wereextracted from the SAS software, version 4.1, and are based on pooled SID numerators (i.e.,the relative frequency of the numerators of the component indicators). This approach is knownas “event weighting,” because it gives equal weight to each event, regardless of how manypatients were evaluated for the occurrence of that event. The PSI Composite weights wereproportionally reallocated into the NHQR/NHDR summary measure weights to account for theexcluded PSIs.Methods for HCUP Data in 2011 NHQR and NHDR D-2December 9, 2011

Calculation of Summary MeasuresEach summary measure was calculated as follows:Where a i corresponds to the weight to the i th QI and X i corresponds to the risk-adjusted rate forthe i th QI.The standard error (SE) of the summary measure is the square-root of the variance:Where a i corresponds to the weight to the i th QI and X i corresponds to the risk-adjusted rate forthe i th QI. The correlations actually had very little effect on the estimated SE for the summarymeasures. For example, in examining mortality for select conditions, the SE was 0.293 if weassume the correlations are zero (i.e., the individual measures are uncorrelated) and the SEwas 0.303 if we assume the correlations are those estimated by the covariance matrix of thestate-level rates, which were in the range of 70 to 85 percent. Therefore, the SEs werecalculated on the assumption that the individual measures were independent of one another,which eliminates the second term on the right-hand side of the formula above.Methods for HCUP Data in 2011 NHQR and NHDR D-3December 9, 2011

APPENDIX E:COMPARISON OF INPATIENT AND EMERGENCY DEPARTMENT USEFOR SELECTED PQIs AND PDIsFor the 2011 NHQR, HCUP data were used to examine national and regional differences ininpatient and emergency department (ED) use for selected AHRQ Prevention Quality Indicators(PQIs) and related Pediatric Quality Indicators (PDIs). Table E-1 in this appendix contains a listof PQIs and PDIs examined.The PQIs are measures of quality associated with processes and outcomes of care thatoccurred in an outpatient or an inpatient setting. The PQIs rely solely on hospital administrativedata and, for this reason, are screens for examining quality that may indicate the need for morein-depth studies. Experts have suggested that using both inpatient and emergency room datamay give a more accurate picture of avoidable visits/admissions for some ambulatory caresensitive conditions which can be identified by certain PQIs and PDIs.Two HCUP databases were used for the analysis:The HCUP Nationwide Emergency Department Sample (NEDS), a nationally stratifiedsample of hospital-based EDs from HCUP States that contribute ED data (28 States inthe 2008 NEDS).The HCUP Nationwide Inpatient Sample (NIS), a nationally stratified sample of hospitalsfrom HCUP States that contribute inpatient data (42 States in the 2008 NIS).The 2008 NEDS contains approximately 28.5 million ED events from 980 hospital-based EDs.The NEDS includes information on ED visits that do not result in an admission (i.e., treat-andreleasevisits and transfers to another hospital) as well as discharge information on patientsinitially seen in the ED and then admitted to the same hospital. For 2008, the NIS containsroughly 8.2 million inpatient discharges from more than 1,000 hospitals. Discharge-levelweights included with the NEDS and NIS are used to produce national estimates.Several steps were taken to prepare the HCUP databases: (1) QI software review andmodification, (2) acquisition of population-based data, (3) general preparation of HCUP data,and (4) identification of statistical methods.1. QI Software Review and Modification. A modification of PQI Version 4.1 was used.The PQIs were developed for use with hospital inpatient discharge data. No guidelinesfor applying the AHRQ QIs to emergency department data were available when thisanalysis began. Some of the events in the NEDS are visits for patients initially seen inthe emergency room and then admitted to the same hospital (an “ED admission”), andsome NEDS events are ED visits that do not result in an inpatient admission (e.g., treatand-releasevisits and transfers to another hospital). About 16 percent of records in the2008 NEDS represent an ED admission. The PQIs rely on the first-listed diagnosis code(DX1) to identify cases with the outcome of interest. For ED admissions, DX1 is theprincipal diagnosis code and reflects the condition established to be chiefly responsiblefor a patients’ admission to the hospital. Unfortunately, principal diagnosis is not clearlydiscernible for ED visits that do not result in admission. Coding instructions for outpatientdata specify that the first-listed diagnosis is supposed to be the "reason for visit," whichis different than the principal diagnosis. Even though DX1 in ED data is not necessarilythe principal diagnosis, using DX1 preserves the concept from the PQI algorithm that theMethods for HCUP Data in 2011 NHQR and NHDR E-1December 9, 2011

first code has higher priority than others. Therefore, this analysis used the first-listeddiagnosis in both the inpatient and ED data analyses.2. Acquisition of Population-Based Data. The next step was to acquire data for thenumerator and denominator populations for the QIs. A QI is a measure of an event thatoccurs in a hospital, requiring a numerator count of the event of interest and adenominator count of the population (within the hospital or within the geographic area) towhich the event relates.For the numerator counts of the PQI or PDI, we used the HCUP NEDS to create nationalestimates of all ED visits, ED visits resulting in admission to the same hospital, and allother types of ED visits. We used the HCUP NIS to create national estimates ofinpatient admissions including those admitted through the ED. For the denominatorcounts, population ZIP-Code-level counts from Claritas (a vendor that compiles andadds value to the U.S. Bureau of Census data) were used for all reporting categories.Claritas uses intra-census methods to estimate household and demographic statistics forgeographic areas (Claritas, Inc., 2008). We also used the Claritas population data forrisk adjustment by age and gender.3. Preparation of HCUP Data. Next, the HCUP NEDS was modified to create an analyticfile consistent with the NIS which is already used for other measures in the NHQR. TheNEDS consists only of hospital-based EDs from community, non-rehabilitation hospitalsand includes discharge weights to the universe of hospital-based ED visits to the U.S. asdefined by the American Hospital Association Annual Survey Database. For missingage and gender data that occurred on a small proportion of discharge records, a “hotdeck” imputation method (which draws donors from strata of similar hospitals andpatients) was used to assign values while preserving the variance within the data.4. Statistical Methods. Age-gender adjustments were made for age and genderdifferences across population subgroups and were based on methods of directstandardization (Fleiss, 1973). Age was categorized into 18 five-year increments.5. Masking Rates for Statistical Reliability, Data Quality, and Confidentiality. PQI andPDI estimates were included in this analysis if they reached a threshold defined by arelative standard error less than 30% and at least 10 unweighted cases in thedenominator. Estimates that did not satisfy these criteria were set to missing.Methods for HCUP Data in 2011 NHQR and NHDR E-2December 9, 2011

Table E-1. List of PQIs and PDIs Used to Examine Differences in Inpatient and ED UsePQI or PDIPQI 1PQI 3PQI 5PQI 7PQI 8PQI 8B*PQI 10PQI 11PQI 12PQI 13PQI 14PQI 15PQI 15B*PQI 16PQI 18*PQI 90PQI 91PQI 92PDI 14PDI 15DescriptionDiabetes with short-term complicationsDiabetes with long-term complicationsChronic obstructive pulmonary diseaseHypertensionCongestive heart failureCongestive heart failure secondary diagnosis with related symptom as firstlisteddiagnosisDehydrationBacterial pneumoniaUrinary tract infectionsAngina without procedureUncontrolled diabetes without complicationsAdult asthmaElderly AsthmaLower extremity amputations among patients with diabetesImmunization-preventable influenzaOverall Prevention Quality Indicator compositeAcute Prevention Quality Indicator compositeChronic Prevention Quality Indicator compositePediatric asthmaPediatric diabetes with short-term complications* Modified or added version of PQI.Methods for HCUP Data in 2011 NHQR and NHDR E-3December 9, 2011

APPENDIX F:READMISSIONS FOR SELECT CHRONIC CONDITIONSFor the 2011 NHQR, HCUP data was used to examine state-level differences in hospitalreadmissions for five chronic conditions – congestive heart failure (CHF), asthma alone, asthmaor chronic obstructive pulmonary disease (COPD), pneumonia, and acute myocardial infarction(AMI).Readmission rates and costs for CHF, asthma, asthma/COPD, pneumonia, and AMI wereexamined for 2008. Fifteen HCUP states provided in their State Inpatient Databases (SID)synthetic de-identified person numbers that allowed an individual patient to be followed withinthe year and across hospitals in the State. States were not identified in the state-level analysis.To prepare the SID for the readmission analysis, the following steps were taken:1. Tracking a patient over time: The HCUP Supplemental Files for Revisit Analysis wereused to track patients across time and hospital settings in the SID. 8 Using thesesupplemental revisit files, a verified person number and information on the number ofdays between inpatient admissions were added to the SID.2. Selection of hospitals: The SID were subset to include only community, nonrehabilitationhospitals. Community hospitals are defined by the American HospitalAssociation as “non-Federal, short-term, general, and other specialty hospitals,excluding hospital units of institutions.” Specialty hospitals included among communityhospitals are obstetrics-gynecology, ear-nose-throat, orthopedic, and pediatricinstitutions. Also included among community hospitals are public hospitals andacademic medical centers. No adjustment was made for community hospitals notrepresented in the SID.3. Transfers: Multiple discharges that represented the transfer of a patient from onehospital to another, or from one unit of a hospital to another, were combined so that thesecond part of each “episode of care” was not counted as a readmission. If for the sameperson, one discharge ended on the same day as a second discharge started, the twodischarge records were combined into a single “transfer” record. The combined transferrecord retained the diagnoses from the second discharge and combined the length ofstay and total hospital charges from the two discharges. The percentage of dischargesthat were transfers in each state ranged from 6 percent to 5.7 percent with an average of3.1 percent.4. Selection of patients: To qualify for the analysis a patient needed to have a verifiedperson identifier (i.e., a ‘visitLink’ value on the Revisit File), be at least 18 years old, andhave at least one index admission between January 1 and November 30, 2008. Adischarge record was considered a index admission if it satisfied the following threecriteria:Target principal diagnosis based on Clinical Classification Software (CCS). CCScategorizes ICD-9-CM diagnoses into 260 clinically meaningful categories. 9 This8 HCUP Supplemental Files for Revisit Analyses. Healthcare Cost and Utilization Project (HCUP). Agencyfor Healthcare Research and Quality, Rockville, MD. www.hcupus.ahrq.gov/toolssoftware/revisit/revisit.jsp.9 HCUP CCS. Healthcare Cost and Utilization Project (HCUP). May 2011. Agency for HealthcareResearch and Quality, Rockville, MD. www.hcup-us.ahrq.gov/toolssoftware/ccs/ccs.jsp.Methods for HCUP Data in 2011 NHQR and NHDR F-1December 9, 2011

"clinical grouper" makes it easier to quickly understand patterns of diagnoses andprocedures. The target CCS for CHF was 108; the target CCS for asthma was128; the target CCS for asthma/COPD was 127 or 128; the target for pneumoniawas 122; and the target for AMI was 100.Discharged alive.Not a combined transfer record.For the CHF analysis, there was a fourth criterion that excluded records with hearttransplant procedures (any ICD-9-CM procedure code in the range 37.51-37.54).All discharges with non-missing patient discharge status were retained for the selectedpatients. A patient was allowed to have multiple index admissions, if they occurred morethan 30 days apart.5. Readmissions: This analysis considered readmissions within 30 days of an indexadmission (defined above). The 30-day window was defined from the end of the firstevent (discharge date) to the beginning of the second event (admission date). Thereadmission could occur at any hospital and was not limited to the hospital of the indexadmission. Readmissions may be discharged in January to December 2008, include apatient discharge status of “died”, and be combined transfer records. Three types ofreadmissions were evaluated for inclusion in the NHQR:Readmissions based on same principal diagnosis: If within 30 days of anindex admission, there was a discharge record for the same person with thesame principal diagnosis CCS, using the CCS categories noted above.Readmissions based on same diagnosis: If within 30 days of an indexadmission, there was a discharge record for the same person with the same CCSas a principal or secondary diagnosis. This type of readmission uses the sameCCS categories identified above.Readmissions for any diagnosis: If within 30 days of an index admission, therewas a discharge record for the same person, regardless of the principal orsecondary diagnosis.For the analysis we looked at both readmission rates and cost using the following methodology:Readmission Rates. Readmission rates were calculated as the percentage of index eventswith at least one readmission by reporting category: age (18-64, 65+), gender,race/ethnicity, and income quartile.Costs. The HCUP databases include information on total hospital charges. Using HCUPhospital-level cost-to-charge ratios based on hospital accounting reports from the Centersfor Medicare and Medicaid Services, 10 total charges are converted to costs. Costs will tendto reflect the actual costs of production, while charges represent what the hospital billed forthe stay. Hospital charges reflect the amount the hospital charged for the entire hospitalstay and does not include professional (physician) fees. The total and average cost was10 HCUP Cost-to-Charge Ratio Files (CCR). Healthcare Cost and Utilization Project (HCUP). August2011. U.S. Agency for Healthcare Research and Quality, Rockville, MD. www.hcupus.ahrq.gov/db/state/costtocharge.jsp.Methods for HCUP Data in 2011 NHQR and NHDR F-2December 9, 2011

calculated for the index events and readmissions by reporting category. The total andaverage cost for readmissions includes all readmissions within 30 days of the index event.Statistical Methods. No risk adjustment was made to the readmission rates or costs.Rates and costs are stratified by two age groups – 18-64 and 65 and older. Estimates wereincluded in this analysis if they reached a threshold defined by a relative standard error lessthan 30% and at least 10 cases in the denominator. Estimates that did not satisfy thesecriteria were set to missing.Methods for HCUP Data in 2011 NHQR and NHDR F-3December 9, 2011

APPENDIX G:STATE-LEVEL BIVARIATE ANALYSIS - STATE PQI RATESRELATED TO OTHER FACTORS, TAKEN ONE AT A TIMEThis appendix shows the factors for which State-specific data could be found to compare to the StatespecificPrevention Quality Indicators (PQI) included in the 2005 (third) NHQR based on 2001discharge data from the SID. State-level PQI rates are shown below with whether or not they werecorrelated with these factors. The results shaded in yellow below denote statistically significantcorrelations. The direction of the relationship and the percent of variation across States explained bythe data element are also shown.(Highlighted text denotes statistically significant results)Key to Conclusions about Associations Found in Appendix A, Tables 1-3, column 3:+ = positive association, statistically significant at p

State-Level Bivariate Analysis of AHRQ Prevention Quality Indicators (PQIs) based on2001 Discharge Data Reported in the 2004 and 2005 NHQRAHRQPreventionQualityIndicators(PQIs)Characteristics of State PopulationsConclusionsAboutAssociationsPercent ofStateVariationExplained(R-square)PQI 1 – AdultAdmissions forShort-termDiabetesComplicationsPQI 3 – AdultAdmissions forLong-termDiabetesComplicationsPrevalence of Obesity in Adults + 43.63%Adult Diabetes Prevalence (Diagnosed) + 25.92%Percent of Population 65 Years and Over ns 0.24%Source of Insurance: Uninsured (as a Percent of thePopulation) + 12.09%Percent of People Below the Poverty Line in the Past 12Months + 33.70%Hospital Bed Supply (Rate/100,000) ns 5.21%Race/Ethnicity: White (as a Percent of the Population) ns 0.38%Race/Ethnicity: Black (as a Percent of the Population) + 46.13%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 1.32%Race/Ethnicity: API (as a Percent of the Population) - 12.52%Race/Ethnicity: Other (as a Percent of the Population) ns 3.96%Race/Ethnicity: Minority (as a Percent of the Population) ns 0.38%Percent of Adult Population at Risk for Heart Disease*** + 15.28%Cardiac Deaths (Rate/100,000) + 55.56%Prevalence of Obesity in Adults + 28.29%Adult Diabetes Prevalence (Diagnosed) + 32.36%Percent of Population 65 Years and Over ns 3.18%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 8.41%Percent of People Below the Poverty Line in the Past 12Months + 26.40%Hospital Bed Supply (Rate/100,000) ns 10.75%Race/Ethnicity: White (as a Percent of the Population) ns 5.69%Race/Ethnicity: Black (as a Percent of the Population) + 28.56%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.45%Race/Ethnicity: API (as a Percent of the Population) ns 1.38%Race/Ethnicity: Other (as a Percent of the Population) ns 8.82%Race/Ethnicity: Minority (as a Percent of the Population) ns 5.69%Methods for HCUP Data in 2011 NHQR and NHDR G-2December 9, 2011

AHRQPreventionQualityIndicators(PQIs)Characteristics of State PopulationsConclusionsAboutAssociationsPercent ofStateVariationExplained(R-square)PQI 4 –PediatricAsthmaAdmissionsPQI 6 –PediatricGastroenteritisAdmissionsPQI 8 – AdultAdmissions forCongestiveHeart FailureAdult Asthma Prevalence ns 1.23%Emphysema Prevalence ns 0.97%Chronic Bronchitis Prevalence ns 5.38%Percent Reporting Cigarette Use in the Past Month + 13.57%HMO Penetration ns 2.65%Percent of People Below the Poverty Line in the Past 12Months ns 7.44%Percent Without Telephone Access + 15.27%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 3.25%Hospital Bed Supply (Rate/100,000) ns 4.22%Air Quality - Particulate Annual Mean ns 0.96%Air Quality - Particulate 24 Hour Average ns 0.64%Air Quality - Ozone 1 Hour Average + 16.99%Race/Ethnicity: White (as a Percent of the Population) ns 5.35%Race/Ethnicity: Black (as a Percent of the Population) + 38.75%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.00%Race/Ethnicity: API (as a Percent of the Population) ns 1.63%Race/Ethnicity: Other (as a Percent of the Population) - 20.45%Race/Ethnicity: Minority (as a Percent of the Population) ns 5.35%HMO Penetration ns 6.86%Percent of People Below the Poverty Line in the Past 12Months + 24.91%Percent of Population that is Foreign-Born ns 1.78%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 2.78%Hospital Bed Supply (Rate/100,000) + 40.32%Race/Ethnicity: White (as a Percent of the Population) ns 0.25%Race/Ethnicity: Black (as a Percent of the Population) + 12.06%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 2.32%Race/Ethnicity: API (as a Percent of the Population) ns 0.37%Race/Ethnicity: Other (as a Percent of the Population) ns 10.38%Race/Ethnicity: Minority (as a Percent of the Population) ns 0.25%Percent of Adult Population at Risk for Heart Disease*** + 41.70%Cardiac Deaths (Rate/100,000) + 76.95%Percent Reporting Cigarette Use in the Past Month + 27.46%Percent Reporting Past Month 'Binge' Alcohol Use ns 2.92%Percent of Population 65 Years and Over ns 8.62%HMO Penetration ns 0.51%Methods for HCUP Data in 2011 NHQR and NHDR G-3December 9, 2011

AHRQPreventionQualityIndicators(PQIs)Characteristics of State PopulationsConclusionsAboutAssociationsPercent ofStateVariationExplained(R-square)PQI 8 – cont’d Percent of People Below the Poverty Line in the Past 12Months + 18.67%Percent of Population that is Foreign-Born ns 2.57%Physician Specialist (Rate/100,000) ns 0.99%Medicare Hospital Payment per Beneficiary + 47.98%Race/Ethnicity: White (as a Percent of the Population) ns 0.33%Race/Ethnicity: Black (as a Percent of the Population) + 34.20%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 3.43%Race/Ethnicity: API (as a Percent of the Population) ns 4.07%Race/Ethnicity: Other (as a Percent of the Population) - 25.83%Race/Ethnicity: Minority (as a Percent of the Population) ns 0.33%PQI 14 – AdultAdmissions forUncontrolledDiabetesWithoutComplicationsPQI 15 – AdultAsthmaAdmissionsPrevalence of Obesity in Adults + 35.10%Adult Diabetes Prevalence (Diagnosed) + 12.38%Percent of Population 65 Years and Over ns 4.62%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 9.16%Percent of People Below the Poverty Line in the Past 12Months + 27.49%Hospital Bed Supply (Rate/100,000) + 25.47%Race/Ethnicity: White (as a Percent of the Population) ns 3.29%Race/Ethnicity: Black (as a Percent of the Population) + 35.54%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.01%Race/Ethnicity: API (as a Percent of the Population) ns 3.94%Race/Ethnicity: Other (as a Percent of the Population) - 14.55%Race/Ethnicity: Minority (as a Percent of the Population) ns 3.29%Adult Asthma Prevalence ns 0.02%Emphysema Prevalence ns 0.25%Chronic Bronchitis Prevalence - 12.23%Percent Reporting Cigarette Use in the Past Month + 12.29%HMO Penetration ns 1.28%Percent of People Below the Poverty Line in the Past 12Months ns 6.86%Percent Without Telephone Access + 15.69%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 0.05%Hospital Bed Supply (Rate/100,000) ns 9.98%Percent of Population 65 Years and Over + 11.24%Air Quality - Particulate Annual Mean ns 2.31%Air Quality - Particulate 24 Hour Average ns 1.64%Air Quality - Ozone 1 Hour Average ns 8.06%Race/Ethnicity: White (as a Percent of the Population) ns 6.46%Methods for HCUP Data in 2011 NHQR and NHDR G-4December 9, 2011

AHRQPreventionQualityIndicators(PQIs)Characteristics of State PopulationsConclusionsAboutAssociationsPercent ofStateVariationExplained(R-square)PQI 15 – cont’d Race/Ethnicity: Black (as a Percent of the Population) + 27.39%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.60%Race/Ethnicity: API (as a Percent of the Population) ns 0.19%Race/Ethnicity: Other (as a Percent of the Population) - 19.90%Race/Ethnicity: Minority (as a Percent of the Population) ns 6.46%PQI 15-65 –Adult AsthmaAdmissions,Age 65+PQI 16 -Diabetes-Related LowerExtremityAmputationsAdult Asthma Prevalence ns 4.23%Emphysema Prevalence ns 3.02%Chronic Bronchitis Prevalence ns 10.90%Percent Reporting Cigarette Use in the Past Month ns 0.60%HMO Penetration ns 0.05%Percent of People Below the Poverty Line in the Past 12Months ns 10.92%Percent Without Telephone Access + 11.62%Source of Insurance: Uninsured (as a Percent of thePop.) ns 3.71%Hospital Bed Supply (Rate/100,000) ns 3.69%Percent of Population 65 Years and Over ns 0.37%Air Quality - Particulate Annual Mean ns 0.39%Air Quality - Particulate 24 Hour Average ns 0.01%Air Quality - Ozone 1 Hour Average ns 2.73%Race/Ethnicity: White (as a Percent of the Population) - 28.23%Race/Ethnicity: Black (as a Percent of the Population) + 21.23%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.79%Race/Ethnicity: API (as a Percent of the Population) + 12.93%Race/Ethnicity: Other (as a Percent of the Population) - 14.27%Race/Ethnicity: Minority (as a Percent of the Population) + 28.23%PQI 14: Adult Admissions for Uncontrolled DiabetesWithout Complications + 30.61%PQI 1: Adult Admissions for Short-term DiabetesComplications + 42.17%PQI 3: Adult Admissions for Long-term DiabetesComplications + 60.26%Percent of Adult Population at Risk for Heart Disease*** + 13.16%Cardiac Deaths (Rate/100,000) + 33.45%Prevalence of Obesity in Adults + 15.71%Adult Diabetes Prevalence (Diagnosed) + 25.88%Percent of Population 65 Years and Over ns 0.00%Source of Insurance: Uninsured (as a Percent of thePop.) ns 3.60%Percent of People Below the Poverty Line in the Past 12Months ns 8.30%Methods for HCUP Data in 2011 NHQR and NHDR G-5December 9, 2011

AHRQPreventionQualityIndicators(PQIs)Characteristics of State PopulationsConclusionsAboutAssociationsPercent ofStateVariationExplained(R-square)PQI 16 – cont’d Hospital Bed Supply (Rate/100,000) ns 0.61%HMO Penetration ns 0.93%Race/Ethnicity: White (as a Percent of the Population) - 14.17%Race/Ethnicity: Black (as a Percent of the Population) + 48.27%Race/Ethnicity: Hispanic (as a Percent of thePopulation) ns 0.54%Race/Ethnicity: API (as a Percent of the Population) ns 0.62%Race/Ethnicity: Other (as a Percent of the Population) - 19.00%Race/Ethnicity: Minority (as a Percent of the Population) + 14.17%(Added) -Immunization-PreventableInfluenzaAdmissionsAmong ElderlyPercent of Adult Population at Risk for Heart Disease*** + 20.95%Cardiac Deaths (Rate/100,000) ns 5.26%Emphysema Prevalence ns 0.99%Chronic Bronchitis Prevalence ns 0.89%Percent Reporting Cigarette Use in the Past Month ns 6.70%Percent Reporting Past Month 'Binge' Alcohol Use ns 9.58%Source of Insurance: Uninsured (as a Percent of thePopulation) ns 1.90%HMO Penetration - 17.28%Percent of People Below the Poverty Line in the Past 12Months ns 11.02%Race/Ethnicity: White (as a Percent of the Population) ns 5.74%Race/Ethnicity: Black (as a Percent of the Population) ns 2.27%Race/Ethnicity: Hispanic (as a Percent of thePopulation) - 14.01%Race/Ethnicity: API (as a Percent of the Population) ns 0.81%Race/Ethnicity: Other (as a Percent of the Population) ns 2.90%Race/Ethnicity: Minority (as a Percent of the Population) ns 5.74%Methods for HCUP Data in 2011 NHQR and NHDR G-6December 9, 2011

APPENDIX H:PATIENT SAFETY IN HOSPITALS IN 2004:TOWARD UNDERSTANDING VARIATION ACROSS STATESBy Susan Raetzman, MSPH, Elizabeth Stranges, MS, Rosanna Coffey, PhD,Marguerite Barrett, MS, Roxanne Andrews, PhD, Ernest Moy, MPH, Jeff Brady, MPHEXECUTIVE SUMMARYBackgroundThe emergence of patient safety as a contemporary health issue has resulted in thedevelopment and use of measures, such as AHRQ’s Patient Safety Indicators (PSI), to trackprogress over time in improving patient safety. National PSI rates have been made availableannually in the National Healthcare Quality Report (NHQR), and state-level PSIs will bereleased in the 2007 edition of the NHQR State Snapshots. The purpose of this analysis is toexplore the extent to which differences across states in PSI scores can be explained and todescribe what might account for those differences. The results are intended to help HCUPPartners and AHRQ respond to inquiries about state-level PSI rate variation, which can besubstantial.Study ApproachThe analysis was performed on the nine State Snapshot PSIs released in the 2007 edition ofthe NHQR State Snapshots; the state PSI rates were obtained by applying AHRQ QualityIndicator software to the HCUP State Inpatient Databases (SID) dataset. 11 The PSIs for up to37 states were compared against 58 state-level factors that can be broadly categorized as (a)state policies that are generally intended to affect the quality of health care delivered in thestate; (b) hospital characteristics; (c) coding practices; and (d) other characteristics such aspopulation and health system characteristics. To the extent possible, we included factors in theexternal environment and factors inside hospitals that were conceptually related to medicalerror, quality improvement, or specific patient safety events. Separate correlations of each PSIand each state-specific factor were conducted (i.e., for each PSI, the analyses statisticallyexamined the relationship between the state rates and a particular state-specific factor).FindingsOverall, we found that only about one in five correlations between the State Snapshot PSIs andpotential explanatory factors were statistically significant. The number of statistically significantassociations for the nine individual PSIs range widely from 0 to 21 out of a possible 60associations, including dummy variables (Table 1). In addition, the nature of the significantPSI/factor associations is mixed in that some have plausible explanations and others do not. Inthe latter case, these may be artifacts of other phenomenon or the result of chance statisticalsignificance, given that nearly 550 correlation analyses were performed (i.e., 9 PSIs times 60independent variables).Although there is no pattern to which associations are statistically significant or their direction atthe individual PSI or factor level, a somewhat different picture is revealed when factors areaggregated. Among factor categories, the most consistent analysis results are those pertaining11 For further detail, see Methods Applying AHRQ Quality Indicators to Healthcare Cost and UtilizationProject (HCUP) Data for the Fifth (2008) National Healthcare Quality Report, HCUP Methods SeriesReport #2008-06.Methods for HCUP Data in 2011 NHQR and NHDR H-1December 9, 2011

to the role of coding in explaining variation in state-level PSIs. Taken together, the codingfactors accounted for one-third (33 percent) of statistically significant associations betweenState Snapshot PSIs and explanatory factors. The findings for this category are strengthenedby the fact that associations were consistently positive in direction (i.e., increases in factorvalues were associated with higher PSI rates). The average number of diagnosis fields filled fordischarges in 2004 yielded the largest number of statistically significant associations,suggesting that higher PSI rates sometimes may reflect greater attention to coding, not justworse health outcomes.DiscussionThe analysis of State Snapshot PSIs identified few state-level factors that showed a consistentpattern of association with the nine state-level PSI rates. We suspect that many of the factorsthat should influence patient safety indicators are too new in development or too remote fromwhere safety problems occur to find strong associations in this state-level analysis. Forexample, state programs that proactively disseminated information to the public or providerswere relatively new in the early 2000s. Also, medical errors and their prevention occur at theprovider, not the state, level. With this simple and aggregated analysis, we are not surprised tofind few conclusive results.As expected, the strongest result was coding practices. In a similar analysis of state-level PSIsand Prevention Quality Indicators (PQIs) conducted in 2003 using 2000 data, one type of codingpractice (use of E codes) had a strong, consistent relationship with PSIs. In the currentanalysis, the average number of diagnosis fields used was an important factor; more fields wereassociated with higher PSI rates. This suggests that states that are leading the way to safermedical practice should expand the number of diagnosis codes reported and collected. Thiswould make room for reporting of medical errors for complex clinical patients who already havenumerous conditions coded on their discharge records. More and better reporting about patientsafety events is essential to learn about and make improvements in the quality of care.One reassuring result is the lack of consistent statistical relations between patient and hospitalcharacteristics and safety measures. This supports our earlier finings and the conventionalwisdom that errors are unintentional, random events that can affect any patient and that allhospitals need to improve safety.Methods for HCUP Data in 2011 NHQR and NHDR H-2December 9, 2011